Allow underscores in hex strings.

Changelog.md

@@ -7,6 +7,7 @@ Breaking changes:
 
 Language Features:
  * Allow global enums and structs.
+ * Allow underscores as delimiters in hex strings.
 
 
 Compiler Features:

docs/types/value-types.rst

@@ -499,7 +499,7 @@ terminate the string literal. Newline only terminates the string literal if it i
 Hexadecimal Literals
 --------------------
 
-Hexadecimal literals are prefixed with the keyword ``hex`` and are enclosed in double or single-quotes (``hex"001122FF"``). Their content must be a hexadecimal string and their value will be the binary representation of those values.
+Hexadecimal literals are prefixed with the keyword ``hex`` and are enclosed in double or single-quotes (``hex"001122FF"``, ``hex'0011_22_FF'``). Their content must be hexadecimal digits which can optionally use a single underscore as separator between byte boundaries. The value of the literal will be the binary representation of the hexadecimal sequence.
 
 Hexadecimal literals behave like :ref:`string literals <string_literals>` and have the same convertibility restrictions.
 

liblangutil/Scanner.cpp

@@ -67,7 +67,7 @@ string langutil::to_string(ScannerError _errorCode)
 	{
 		case ScannerError::NoError: return "No error.";
 		case ScannerError::IllegalToken: return "Invalid token.";
-		case ScannerError::IllegalHexString: return "Expected even number of hex-nibbles within double-quotes.";
+		case ScannerError::IllegalHexString: return "Expected even number of hex-nibbles.";
 		case ScannerError::IllegalHexDigit: return "Hexadecimal digit missing or invalid.";
 		case ScannerError::IllegalCommentTerminator: return "Expected multi-line comment-terminator.";
 		case ScannerError::IllegalEscapeSequence: return "Invalid escape sequence.";
@@ -759,13 +759,25 @@ Token Scanner::scanHexString()
 	char const quote = m_char;
 	advance();  // consume quote
 	LiteralScope literal(this, LITERAL_TYPE_STRING);
+	bool allowUnderscore = false;
 	while (m_char != quote && !isSourcePastEndOfInput())
 	{
 		char c = m_char;
-		if (!scanHexByte(c))
-			// can only return false if hex-byte is incomplete (only one hex digit instead of two)
+
+		if (scanHexByte(c))
+		{
+			addLiteralChar(c);
+			allowUnderscore = true;
+		}
+		else if (c == '_')
+		{
+			advance();
+			if (!allowUnderscore || m_char == quote)
+				return setError(ScannerError::IllegalNumberSeparator);
+			allowUnderscore = false;
+		}
+		else
 			return setError(ScannerError::IllegalHexString);
-		addLiteralChar(c);
 	}
 
 	if (m_char != quote)

test/libsolidity/semanticTests/literals/hex_string_with_underscore.sol

@@ -0,0 +1,7 @@
+contract C {
+    function f() public pure returns(bytes memory) {
+        return hex"12_34_5678_9A";
+    }
+}
+// ----
+// f() -> 32, 5, left(0x123456789A)

test/libsolidity/syntaxTests/literals/hex_string_duplicate_underscore.sol

@@ -0,0 +1,7 @@
+contract C {
+    function f() public pure {
+        hex"12__34";
+    }
+}
+// ----
+// ParserError: (52-60): Invalid use of number separator '_'.

test/libsolidity/syntaxTests/literals/hex_string_leading_underscore.sol

@@ -0,0 +1,7 @@
+contract C {
+    function f() public pure {
+        hex"_1234";
+    }
+}
+// ----
+// ParserError: (52-57): Invalid use of number separator '_'.

test/libsolidity/syntaxTests/literals/hex_string_misaligned_underscore.sol

@@ -0,0 +1,7 @@
+contract C {
+    function f() public pure {
+        hex"1_234";
+    }
+}
+// ----
+// ParserError: (52-56): Expected even number of hex-nibbles.

test/libsolidity/syntaxTests/literals/hex_string_trailing_underscore.sol

@@ -0,0 +1,7 @@
+contract C {
+    function f() public pure {
+        hex"1234_";
+    }
+}
+// ----
+// ParserError: (52-61): Invalid use of number separator '_'.

test/libsolidity/syntaxTests/literals/hex_string_underscores_valid.sol

@@ -0,0 +1,3 @@
+contract C {
+    bytes constant c = hex"12_3456_789012";
+}